Origins of the Internet

One of the first ideas to get out of your mind (if it is there) is that the Internet is a monolithic creation that came about at a single point in time and at a single location. The Internet is a conglomerate of overlapping and mutually reinforcing technologies from computer science, data storage and retrieval sciences (the once lowly data processing), and communications that have been developed (and are still being developed) over the past half century.

From inception in the Department of Defense (DoD) and the research centers of premier universities, the Internet has been radically transformed time and time again as new hardware and software technologies that are continually developed stream into the global information matrix, mature, and are made obsolete in a seemingly endless evolution. Rather than a single, static entity, the Internet is a living, pulsating, vibrant, multifaceted creation that is still rapidly evolving and changing our world in ways unimagined. Just like our grandparents and great-grandparents at the onset of the Age of the Automobile, we have not fathomed the societal changes the Internet will bring about.

Perhaps by watching our children in their eager embrace of cell phones, portable computers, iPods, instant messaging (IM), MySpace, and YouTube, our generation may gain insight into the near-term transformations of our society that is, if we are not already too mentally numbed by earlier society-shattering discoveries to take account of them. The Internet had its origins more than 40 years ago in 1962 during ongoing research into the possibility of using packets for communication as opposed to switched circuits. One of the early researchers (and most successful proponents of packet-switched circuits) was Leonard Kleinrock of the Massachusetts Institute of Technology (MIT). The concept of a network, however, over which to send the packets of information belongs to J. C. R. Licklider, also of MIT, who presciently wrote about the Galactic Network that could be used by anyone to access information contained in any connected computer.

Emergence of packet switching and ARPANET

All communication takes place from sender to receiver over a communications channel. Ever since the telephone era began, communications have taken place over copper lines, called circuits. You may recall seeing early twentieth or late nineteenth century photos of women in long skirts plugging lines into a switchboard. They were actually connecting a sender and a receiver using the phone lines. When one line (or channel) was being used, no other person or machine could use it. So, if there were three lines connecting Wichita to Omaha, and four people wanted to use them, one person was out of luck until one of the other three people decided to hang up. Then, for example, a person in Wichita could make a call. Some of these calls could be very long (during which time no one else could contact anyone else on that line) and costly (the funds from which went into improving services and research into newer technologies). One of the motivating forces behind the development of packet-switching technology was to make more efficient (and, therefore, cheaper) use of the communications infrastructure. The idea was to take conversations from several different people, put these conversations into separate boxes, address them, and then send those boxes across a single circuit at the same time.

The advantage was efficiency because you could use one line to send lots of conversations across the same line at practically the same moment. Conversations from Cousin Sue, Preacher Paulsen, Freda the Librarian, Farmer Phil, and dozens of others, for example, could all be put in their boxes and shipped across the same single line to Omaha at the same time. Packet switching meant that only one line was needed to carry multiple conversations, whereas circuit switching required one entire line to be used for each conversation. What this meant to the nascent Internet Age was that there was the potential to send greater amounts of data quicker and much more cheaply.

In the mid-to-late 1960s, research continued into a packet-switched network that would allow access to remote computers on multiple networks for data processing and the transfer of data files. Now, this is not to say that you could not access separate networks already. You could. It was just very clumsy. For each network to which you wanted to be connected, you would need a different computer that was physically connected to that one network. This is the same as you having to carry around a different phone for each of your friends you might wish to give you a ring.

In 1964, initial funding for what would become such a system, the DoD’s Advanced Research Projects Agency (ARPA) network (ARPANET), was provided to research ways to network separate systems together. Under the directorship of Larry Roberts, what eventually became (at the end of the decade) ARPANET consisted of four nodes connected to each other via phone lines leased from the telephone company.

Located at the University of California at Los Angeles (UCLA), University of California at Santa Barbara (UCSB), Stanford Research Institute (SRI), and the University of Utah, the four nodes were connected by Interface Message Processors (IMPs). These were the world’s first routers. The creation of ARPANET in 1969, also the year of the first landing on the moon, passed without much public notice.

The simple and revolutionary IMP provided a common interface for all the disparate machines operating on several different operating systems to communicate across a common channel. This is the equivalent of having a universal language translator today. The protocol to route data from node to node (or host to host) was called the Network Control Protocol (NCP). Developed by the Networking Work Group, NCP was made obsolete only in 1983 when it was replaced by the Transmission Control Protocol/Internet Protocol (TCP/IP).

How a router routes

Arouter is actually a small computer used to route communications from source to destination. How it does this is rather complex in actuality, but conceptually it’s simple. An analogy often used to illustrate the concept is that of a post office delivering a letter. When you use ‘‘snail mail,’’ you place the letter in an envelope and write on the envelope two addresses: one is yours (the sender), and the other is address of the place and person you wish to send it to (the destination or receiver). You then put the envelope in the mailbox for the postal carrier to pick up (or you drop it off at the local post office). After addressing and dropping it off, you are finished with it. (If you can see parallels to this and e-mailing by typing in an address, say, myfriend@neighborhood.org, and then clicking the Send button, good for you. You know where we are going with this.) However, your local post office or mailbox is not the final destination of the envelope. It is the first of many intermediary stops.

A letter usually, goes from a local post office in a neighborhood district or small town that feeds into a larger regional office that accepts mail from smaller surrounding stations. At this central post office (CPO), the mail is sorted and packed into bags. These sacks of mail are then sent to other regional and/or local offices. Eventually, after passing through these regional intermediaries, the letter is sent to the local destination office and, from there, delivered to the recipient.

Now, a router functions the same way. It accepts incoming mail (data packets) and looks at the address of the destination. Then it looks at its list of linked router addresses it keeps in its routing table. The router selects the best route to the destination and forwards the mail to it. That intermediary router in turn looks at its routing tables to select the correct port to send the mail out to reach its ultimate destination, and sends it on. The number of intermediate stops (or hops, as they are called) varies depending upon the location of the recipient on the Internet (just as the number of intermediate stations a letter goes to depends upon its geographic location in relation to the sender). If you were to send a letter to Chicago from Fairfax, Virginia, it might go to intermediate post offices in Knoxville, Lexington, Indianapolis, and then on to Chicago a total of four hops, using networking lingo. Just mentally substitute routers for the local and intermediate post offices, and you know how a router works and what a router is.

Just like the post office, a router reads addresses and, based upon the destination address, forwards the packets and selects the most efficient path from sender to receiver.